Because the applications requiring constantly evolving vision systems are so diverse, imaging technology must be adaptable to specific market requirements. For example, low-cost home camera systems require simple, limited functionality, while commercial cameras require more sophisticated features such as high resolution, noise reduction, and enhanced security.

In line with this, Arm announced the launch of its new Mali-C55 Image Signal Processor (ISP) on the 9th. The new ISP is Arm's smallest and most configurable product to date, and is reportedly attracting interest from partners, including Renesas, the first public licensee.

Semiconductor partners and OEMs need the freedom to add or remove features based on application requirements, which is why Arm designed the Mali-C55 with a highly configurable ISP. The Mali-C55 not only supports multiple cameras and integrates with machine learning accelerators, but also includes industry-standard AXI and AHB interfaces for easy integration with Cortex-A or Cortex-M-based SoCs.

The Mali-C55 delivers upgraded image quality and operates in a variety of lighting and weather conditions. It is designed to deliver maximum performance and functionality in area- and power-constrained applications, making it ideal for smart camera and edge AI vision use cases.

In use cases, the Mali-C55 enables surveillance and security cameras to accurately recognize license plates of vehicles moving at speeds up to 75 mph (120.7 km/h), enabling them to detect critical details. Home cameras and security systems can capture high-resolution images both indoors and outdoors, and smart home hubs can even feature advanced features like secure visual unlocking.

Arm said the Mali-C55, which supports multi-camera capabilities for up to eight individual inputs, image resolutions up to 8K, and image sizes up to 48 megapixels (MP), offers advantages in image quality, throughput, power consumption, and semiconductor area.

Building on the class-leading Mali-C52 ISP, the Mali-C55 features improved tone mapping and spatial noise reduction, enhanced support for high dynamic range (HDR) sensors, and seamless integration with machine learning accelerators that utilize neural networks for various noise reduction techniques. Combining multiple Mali-C55 ISPs also enables larger image sizes for applications that require more than 48 MP, such as video conferencing.

Silicon footprint and cost are important factors in embedded and IoT vision applications. With the Mali-C55, Arm delivers improved functionality in nearly half the silicon area of previous generations, significantly reducing power consumption and extending battery life, which also reduces device costs.

As machine learning moves closer to the edge, more ISPs can be integrated into systems-on-chip (SoCs) to leverage advanced image processing. Arm is leveraging the integration of the Mali-C55 with machine learning accelerators to deliver a new level of on-device processing for devices requiring high-quality vision systems. This is because the ISP's output can be sent directly to the machine learning accelerator, reducing the amount of data transferred from the device to the cloud without having to compromise inference, reducing costs and processing times.

“ISPs are one of the most critical information-generating devices supporting a wide range of IoT vision system applications, including smart cameras and drones for commercial, industrial, or home use,” said Mohamed Awad, vice president and general manager of the IoT and Embedded Business Line at Arm. “As demand for high-quality image processing in future devices grows, Arm’s ISP technology roadmap is an area that requires continued investment.”